Latent cellular analysis robustly reveals subtle diversity in large-scale single-cell RNA-seq data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables detailed cell variation analysis.
• scRNA-seq data presents challenges: sparsity, batch effects, and scalability.
• Existing algorithms struggle with complex scRNA-seq data, impacting subpopulation identification.

Purpose of the Study:

• To develop a robust machine learning pipeline for scRNA-seq data analysis.
• To address limitations of current algorithms in identifying cell heterogeneity.
• To create a scalable solution for large-scale scRNA-seq datasets.

Main Methods:

• Developed Latent Cellular Analysis (LCA), a machine learning pipeline.
• Combined cosine-similarity measurement of latent cellular states with graph-based clustering.
• Implemented heuristic solutions for population inference, dimension reduction, and feature selection.

Main Results:

• LCA demonstrated robustness and accuracy compared to state-of-the-art methods.
• Successfully applied to large-scale real and simulated scRNA-seq data.
• LCA's scalability addresses challenges posed by increasing sample sizes.

Conclusions:

• LCA offers a powerful and scalable analytical approach for scRNA-seq data.
• The pipeline effectively identifies biologically meaningful cell subpopulations.
• LCA overcomes key computational hurdles in single-cell data analysis.